【Objective】 To establish a highly sensitive detection method of platelet HPA-3 and HPA-15 genotyping by droplet digital PCR (ddPCR), and to explore the feasibility of applying it to the detection of human platelet antigen (HPA) compatibility in maternal peripheral blood fetal free DNA. 【Methods】 For SNP mutation sites of HPA-3 and HPA-15, specific primers and MGB probes were designed, and amplification conditions such as annealing temperature and primer concentration of ddPCR were optimized to establish the optimal reaction system and clarify the test procedures. The methodological performance of the assay was evaluated, including specificity, sensitivity, repeatability and stability. ddPCR was used to detect 67 clinical blood samples, and the allele typing results were compared with the gene sequencing results. The fetal free DNA HPA antigen of 52 maternal peripheral blood samples was detected. 【Results】 The ddPCR method for detecting platelet HPA-3 and HPA-15 showed good specificity of primers and probes. The optimal annealing temperatures for HPA-3 and HPA-15 were 61.6℃ and 60.2℃, respectively. The optimal concentrations of primers were 900 nM and 700 nM respectively. The final concentration of the probe was 250 nM. The quantitative detection range of copy number was 2 to 20 000 copies, with lower limit of detection of 0.1 copies/μL, and the linearity is good. In low copy number samples, the intra - and inter batch coefficient of variation (CV) of actual detection values for HPA-3 and HPA-15 were both lower than 5%. The detection results of HPA-3 and HPA-15 genotypes of 67 blood samples were consistent with the gene sequencing results, and its application in fetomaternal platelet HPA-3, HPA-15 genotype detection met expectations. 【Conclusion】 The HPA-3 and HPA-15 ddPCR detection system constructed in this study has high accuracy, good repeatability, stability and sensitivity, and can be applied to the establishment of platelet HPA-3 and HPA-15 genotype donor pool, gene matching and fetomaternal platelet compatibility detection.